Skip to main content
NCBI home page
VA Author Manuscripts logoLink to VA Author Manuscripts
. Author manuscript; available in PMC: 2026 Feb 20.
Published in final edited form as: J Affect Disord. 2025 Feb 17;377:124–133. doi: 10.1016/j.jad.2025.02.045

Functional brain connectivity of the salience network in alcohol use and anxiety disorders

Dhruv M Patel a,b, Guillermo F Poblete c, Alexandra Castellanos d,e, Ramiro Salas a,c,d,e,*
PMCID: PMC12917952  NIHMSID: NIHMS2145319  PMID: 39971011

Abstract

The interplay between alcohol use disorder (AUD) and anxiety disorders (ANX) is well-documented, yet the underlying neurobiological mechanisms remain elusive. This study aims to elucidate these mechanisms by examining the resting-state functional connectivity (RSFC) within the salience network and to the amygdala, both implicated in alcohol and anxiety disorders. We analyzed data from 264 inpatient participants culled from a wider group of 518 inpatients at The Menninger Clinic in Houston, TX, categorized into four groups (n = 66 each) based on DSM-IV diagnoses: AUD without ANX (AUD), ANX without AUD (ANX), concurrent AUD and ANX (BOTH), and neither (NEITHER). Our findings reveal significant RSFC differences, particularly between the right supramarginal gyrus (SMG) and 1) right rostral prefrontal cortex (RPFC) (corrected p = 0.029; RSFC significantly higher in NEITHER than in BOTH), and 2) left supramarginal gyrus (SMG) (corrected p = 0.016; RSFC significantly higher in AUD and NEITHER than in BOTH). Furthermore, correlations with a clinical measure for alcohol use (World Health Organization Alcohol, Smoking and Substance Involvement Screening Test; WHO ASSIST) indicated significant relationships: WHO ASSIST alcohol scores negatively correlated with right SMG to right RPFC RSFC (r = −0.14, p = 0.02) and positively correlated with the interhemispheric SMG RSFC (r = 0.17, p = 0.006). This research enhances our understanding of the complex neurobiological interconnections between alcohol use and anxiety disorders, suggesting a disrupted neural architecture that may underpin the behavioral manifestations observed in these highly comorbid conditions.

Keywords: Alcohol use disorder, Anxiety, Salience network, MRI, Supramarginal gyrus, Rostral prefrontal cortex

1. Introduction

The interaction between alcohol use disorder (AUD) and anxiety disorders (ANX) involves shared and overlapping neural circuits (Anker, 2019). Clinically, the co-occurrence of AUD and ANX often leads to a more severe course for each, characterized by greater impairment and a higher risk of relapse (Driessen et al., 2001; Kushner et al., 2005). Prospective studies further highlight a bidirectional relationship, with anxiety disorders increasing the likelihood of developing AUD and vice versa (Kushner et al., 1999). This comorbidity complicates treatment, as maladaptive coping strategies for one may exacerbate symptoms of the other. For example, the use of alcohol to self-medicate anxiety symptoms can lead to dependency and worsen anxiety over time, a pattern described by the self-medication hypothesis (Khantzian, 1997). Moreover, individuals with comorbid anxiety and AUD are more likely to relapse as a way to cope with emotional distress, underscoring the maladaptive nature of this behavior (LaBounty et al., 1992). Addressing comorbid anxiety disorders during treatment, however, has been shown to improve alcohol-related outcomes, reinforcing the importance of integrated therapeutic approaches (Kushner et al., 2000).

AUD involves complex neurobiological pathways and brain adaptations that are crucial for understanding its development and persistence. Chronic alcohol consumption significantly impacts neurotransmitter systems such as serotonin, dopamine, gamma-aminobutyric acid, glutamate, acetylcholine, and opioid systems. These neurotransmitters are essential for mood, reward, and behavioral regulation, which are markedly disrupted in AUD (Yang et al., 2022). Moreover, alcohol influences various brain circuits, especially those involved in executive functions and emotional regulation (Yang et al., 2022). Functional magnetic resonance imaging (fMRI) studies have further elucidated the specific brain regions activated by alcohol-related cues and stress in individuals with AUD (Fukushima et al., 2020; Gorka et al., 2013; Grodin et al., 2019). ANX encompasses a range of conditions where pervasive and persistent anxiety negatively affects an individual’s daily functioning. These disorders include generalized anxiety disorder, panic disorder, social anxiety disorder, and specific phobias. Each of these conditions involve excessive fear or worry, but they differ in the objects or situations that trigger anxiety. Anxiety disorders are highly prevalent and can significantly impair social, occupational, and other important areas of functioning (Wilmer et al., 2021). Functional magnetic resonance imaging (fMRI) has provided significant insights into the neural correlates of anxiety disorders (Won et al., 2020) (Liu et al., 2015; Khoury et al., 2018). The Salience Network (SN) is composed of interconnected brain regions, with the anterior insula and anterior cingulate cortex identified as crucial nodes (Menon et al., 2010). Regions such as the inferior parietal cortex and lateral prefrontal cortex have also been linked to the SN (Gordon et al., 2017). Additionally, subcortical structures, including the amygdala, play significant roles in this network (Uddin, 2015). Despite these associations, there is ongoing debate about the organization of the SN (Uddin et al., 2019). For instance, the salience network has been referred to as the cingulo-opercular network or ventral attention network (Dosenbach et al., 2007; Fox et al., 2006; Yeo et al., 2015). In our study, we used the CONN toolbox, which provides a set of prebuilt regions of interest (ROIs) and networks for analyzing functional connectivity. This choice allowed us to adopt a replicable approach for exploring the SN connectivity patterns relevant to AUD and ANX.

The SN, central to processing environmental stimuli and internal states, plays a critical role in addiction and anxiety disorders (Cushnie et al., 2023). The SN is involved in detecting and filtering salient stimuli and is extensively connected with other brain networks (Menon et al., 2010; Peters et al., 2016). The SN is linked to cognitive and attention control, behavioral modulation, and sensory input processing (Peters et al., 2016). Alterations in SN activity are considered to have clinical effects (Uddin, 2016). Dysregulation in this network can lead to an exaggerated focus on negative or threatening stimuli, enhancing the risk for both developing and perpetuating AUD and anxiety disorders (Schimmelpfennig et al., 2023). Studies have shown that targeting the SN by modulating its activity might be effective in treating AUD, suggesting a promising area for therapeutic interventions (Padula et al., 2022). The SN is composed of several regions including the anterior cingulate cortex (ACC), insula, supramarginal gyrus (SMG), and rostral prefrontal cortex (RPFC). The ACC is known for its role in cognitive control and emotion regulation (Seeley et al., 2007). The ACC is involved in craving responses and conflict monitoring, particularly when individuals are exposed to alcohol-related cues. It also plays a critical role in anxiety, where its heightened activity correlates with the excessive worry and fear characteristic of anxiety disorders (Bush et al., 2000; Paulus et al., 2005). The insula integrates interoceptive awareness with emotional salience to influence decision-making (Menon et al., 2010) and is heavily implicated in both AUD and anxiety disorders. For AUD, the insula is critical in the subjective experience of craving and the urge to drink, while in anxiety, it processes fear and anxiety-related bodily sensations. Dysfunction in the insula can exacerbate the severity of both conditions (Naqvi et al., 2010; Simmons et al., 2006). The SMG plays a role in the SN by integrating sensory information, which is essential for empathy and social cognition (Bzdok et al., 2013). The SMG’s role in empathy and social processing can impact social anxiety and behaviors seen in AUD. Disruptions in SMG activity might affect social judgment, potentially influencing social anxiety symptoms and social contexts that trigger alcohol use (Silani et al., 2013). RSFC between the bilateral SMG and the left amygdala increased with successful anxiety treatment, while RSFC between the bilateral SMG and the right anterior insula of the SN is positively associated with improvement in anxiety self-report (Munjuluri et al., 2020). The RPFC is involved in various higher cognitive functions, including goal setting and the anticipation of future events (Gilbert et al., 2006). In anxiety disorders, the RPFC is involved in excessive rumination and worry. Its dysfunction can lead to poor regulation of attention between internally and externally directed information, a common issue in ANX. In AUD, this might relate to the impaired decision-making and prioritization of alcohol use over other life demands (Gilbert et al., 2006; Moss et al., 2015).

The amygdala, a brain region associated with emotional processing, particularly fear, is often found to be hyperactive in individuals with anxiety disorders. This increased activity is correlated with heightened emotional responses to perceived threats and is a hallmark feature in disorders such as generalized anxiety disorder and panic disorder (Won et al., 2020). Many anxiety disorders are characterized by heightened activity and reactivity of the amygdala (Shin et al., 2010). In individuals with alcohol dependence, the amygdala often exhibits reduced volume, and these structural changes are associated with intensified cravings compared to healthy individuals (Wrase et al., 2008). This dual role highlights the amygdala’s importance as a central node in the overlapping neural mechanisms underlying these comorbid conditions.

This study investigated the resting state functional connectivity (RSFC) within the SN and to the amygdala in a hypothesis-based analysis in psychiatric inpatients with AUD and/or anxiety disorders, and psychiatric controls. This work addresses a critical research gap in understanding how SN alterations manifest in individuals with comorbid AUD and ANX, known to result in more severe clinical outcomes than either condition alone. We hypothesized that RSFC within the SN and to amygdala would be significantly different between the four patient groups of interest. Additionally, we examined whether the severity of AUD and ANX symptoms experienced by inpatients, measured by the WHO ASSIST and GAD-7, correlated with RSFC alterations. Understanding whether RSFC alterations are additive, synergistic, or antagonistic is crucial for predicting clinical outcomes and optimizing treatment strategies, as this insight could inform the development of targeted therapies for AUD-ANX comorbidity.

2. Materials and methods

2.1. Participants

Psychiatric patients (N = 518) were recruited from The Menninger Clinic in Houston, Texas as a part of the McNair Initiative for Neuroscience Discovery – Menninger/Baylor (MIND-MB) study (Ambrosi et al., 2024; Gosnell et al., 2020; Poblete et al., 2022; Rufino et al., 2024). MIND-MB was a five-year neuroscience research initiative (2012 to 2017) aimed at creating a psychiatric inpatient dataset to enhance understanding of mental disease progression and treatment outcomes. This study received approval from the Institutional Review Board at Baylor College of Medicine. Most inpatients exhibited co-occurring psychiatric illnesses, including depressive, anxiety, substance use, and personality disorders. Participants with comorbid disorders were included, to reflect the pervasive comorbidity in psychiatry, including between anxiety disorders and SUD (Conway et al., 2006) and the need to investigate ecologically valid samples in neuroimaging studies (Greene et al., 2016). Participants provided written informed consent before inclusion. Participants were neither consuming substances nor in acute withdrawal during data collection. All patients were medicated at time of assessment.

2.2. Clinical measures

The Structured Clinical Interview for DSM-IV Axis I and II disorders was employed to establish psychiatric diagnoses shortly after the patients’ admission to the clinic. Additional assessments were conducted to evaluate various psychiatric characteristics, including anxiety using the Generalized Anxiety Disorder 7-item (GAD-7) and alcohol use with the World Health Organization Alcohol, Smoking and Substance Involvement Screening test (WHO ASSIST (Humeniuk et al., 2008)). The GAD-7 (Plummer et al., 2016) is a 7-item scale that assesses generalized anxiety disorder severity on a 0–21 point scale, based on symptom frequency over the past two weeks. WHO ASSIST assesses patients’ substance use and related problems over the past 3 months and uses the following scale for alcohol use: low (0–10), moderate (11–26), and high (27+).

2.3. Cohort creation

For the current analysis, a subset of 264 participants was selected from the original cohort of 518 inpatients. This culling process ensured balanced representation across the four diagnostic groups. Patients were classified based on DSM-IV diagnoses: solely AUD, solely ANX, concurrent AUD and ANX (BOTH), and neither (NEITHER). The AUD group encompassed diagnoses of alcohol dependence or alcohol abuse, reflecting the new unified diagnosis of alcohol use disorder per the DSM-5 (First et al., 2002). The ANX group included diagnoses of general anxiety disorders, panic disorder (with or without agoraphobia), agoraphobia, specific phobia, and unspecified anxiety disorders. Covariates of age, sex, and other specific psychiatric diagnoses were normalized prior to matching. Group matching using an Euclidean distance minimization algorithm based on our previously described method (Gosnell et al., 2020) was performed. This algorithm matched groups in a one-to-one manner, calculating the Euclidean distance between each pair of groups such that the sum of all paired distances was minimized. Covariate balance in the matched sample was verified using chi-squared tests for categorical variables and ANOVAs for continuous variables, with p > 0.05 accepted for each variable to be matched between each pair of groups.

2.4. Brain imaging

Inpatients were scanned with a 3 T Siemens Trio MR scanner in the Core for Advanced MR Imaging at Baylor College of Medicine, as close to admission as possible. A structural MPRAGE sequence (1 mm isotropic voxels, TR = 1200 ms, TE = 2.66 ms, flip angle = 12°, field of view (FOV) = 245 mm, total acquisition time = 4.5 min) was acquired, followed by a resting state echo-planar imaging (EPI) scan (3.4 × 3.4 × 4 mm, TR = 2000 ms, TE = 30 ms, flip angle = 90°, FOV = 220 mm, total acquisition time = 5 min). The resting fMRI data were preprocessed with CONN v15.b (Whitfield-Gabrieli et al., 2012). Functional images were down-sampled to 3-mm isotropic voxels and preprocessed similarly to our previous work (Ambrosi et al., 2019; Curtis et al., 2019; Curtis et al., 2017), including functional realignment, slice-timing correction, structural normalization to the MNI template, functional normalization and smoothing with an 8-mm full width at half maximum Gaussian smoothing kernel. The Artifact Detection Toolbox (built-in CONN) was used for outlier detection and scrubbing to create confound regressors for motion estimates using default parameters (framewise displacement above 2 mm translation and 2° rotation in any direction, global threshold: 9 standard-deviations above or below the mean). With the anatomical component correction (aCompCor) we identified the top five principal components for each patient that are associated with segmented white matter and cerebrospinal fluid (Behzadi et al., 2007) and incorporated those components as potential confounds along with realignment parameters and nuisance regressors from the ART-based scrubbing. Functional images were temporally band-pass filtered (0.008–0.09 Hz). CONN automatically excluded excessive motion data (Conn default intermediate settings were used), but no entire participants were removed.

2.5. Functional connectivity

Whole-brain regions of interest (ROIs) were selected from the Automated Anatomical Labeling (AAL) atlas for the resting state functional connectivity analysis (Tzourio-Mazoyer et al., 2002). The mean time series in each ROI and Pearson correlations of the mean time series between each pair of ROIs was computed. The SN was defined with the following ROIs: anterior cingulate cortex (ACC; 0, 22, 35), left anterior insula (L AInsula; −44, 13, 1), right anterior insula (R AInsula; 47, 14, 0), left rostral prefrontal cortex (L RPFC; −32, 45, 27), right rostral prefrontal cortex (R RPFC; 32, 46, 27), left supramarginal gyrus (L SMG; −60, −39, −31), and right supramarginal gyrus (R SMG; 62, −35, 32). The amygdala was defined with the default Conn atlas settings.

2.6. Statistical analysis

One-way ANOVAs were performed using the RSFC values between the diagnostic groups. To control for multiple comparisons in our analysis, we applied the Hochberg procedure, a step-up approach to adjust p-values while maintaining the Familywise Error Rate (FWER) at a significance level of α = 0.05 (Vickerstaff et al., 2019). When an interaction was significant, post hoc analysis was performed through Tukey’s Honestly Significant Difference (HSD) tests, to adjust for multiple comparisons, pinpointing the specific pairs of patient groups that contributed to the significance of the interactions observed. This approach identified differences in brain connectivity relevant to the distinct clinical profiles.

For RSFCs showing significant differences between cohorts (corrected p < 0.05), two-way ANOVAs were conducted to investigate potential interaction effects between AUD and ANX diagnoses. This analysis enabled the identification of both main effects and potential interaction effects, to clarify the nature of their combined influence on connectivity. To perform correlations between clinical dimensional data and RSFC, we used WHO ASSIST and GAD-7 scores. We consider these correlations an exploratory analysis, as no multiple comparison corrections were performed.

3. Results

Fig. 1 and Table 1 show the demographic and clinical characteristics of the four patient groups (n = 66 each) in the matched study sample (N = 264). The matching algorithm yielded comparable groups, with no significant differences in any of the studied variables. The average age of the original cohort of MIND-MB participants was 31.0 ± 12.3 years, with 56.0 % being male (290/518) and 87.3 % identifying as Caucasian. Medication data was unavailable for six participants in total, with two from each group: NEITHER, AUD, and BOTH.

Fig. 1.

Fig. 1.

Open in a new tab

Clinical characteristics of patients presented as psychiatric disorders from the DSM-IV to show the prevalence of disorders within the sample. DYSTHYMIC: Dysthymic disorder; PSYCHOTICNOS: Unspecified psychosis; BIPOLARI: bipolar I disorder, BIPOLARIFEA: bipolar I disorder with psychotic features, BIPOLARII: bipolar II disorder, MDDSINGLE: major depressive disorder, single episode; MDDRECUR: major depressive disorder, recurrent; DEPRESS: depressive disorder; OCD: obsessive compulsive disorder; PTSD: posttraumatic stress disorders; BODYDYS: body dysmorphic disorder; BULIMIA: bulimia nervosa; ANOREXIA: anorexia nervosa; BINGED: binge eating disorder; EDNOS: eating disorder, not otherwise specified. No disorder showed statistically significant prevalence differences between the four experimental groups.

Table 1.

Demographic and clinical data of patient groups.

NEITHER ANX AUD BOTH χ2(3) or F (3,260) p
Gender (Nmale/Nfemale) 40/26 40/26 32/34 41/25 3.28 0.35
Age: mean (SD) 28.6 (10) 28.0 (9.7) 28.8 (11.2) 29.7 (10.0) 0.33 0.80
ASSIST: mean (SD) 5.6 (4.8) 6.7 (7.4) 24.5 (10.6) 24.3 (10.5) 95.4 <0.001
GAD-7: mean (SD) 10.4 (6.4) 11.5 (5.6) 10.9 (6.0) 12.2 (5.5) 1.159 0.32
Antidepressants: N (%) 39 (59.1) 52 (78.8) 49 (74.2) 49 (74.2) 6.70 0.08
Mood Stabilizers: N (%) 31 (47.0) 34 (51.5) 26 (31.4) 33 (50.0) 2.05 0.56
Antipsychotics: N (%) 8 (12.1) 8 (12.1) 13 (19.7) 11 (16.7) 2.29 0.52
Anxiolytics and Sedatives: N (%) 20 (30.3) 27 (40.9) 20 (30.3) 20 (30.3) 2.05 0.56
Stimulants: N (%) 19 (28.8) 15 (22.7) 12 (18.2) 11 (16.7) 3.47 0.33
Pain Management: N (%) 16 (24.2) 14 (21.2) 13 (19.7) 17 (25.8) 0.97 0.81
Open in a new tab

Table 2 shows the RSFC for all pairs of regions studied. Significantly different RSFC was noted between the right SMG and 1) the right RPFC (p < 0.001; corrected p = 0.029; highlighted in blue), and 2) the left SMG (p < 0.001; corrected p = 0.016; highlighted in red). Post hoc analysis using Tukey’s HSD (Fig. 2) revealed the NEITHER cohort to have a significantly different RSFC between the right SMG and right RPFC than the BOTH cohort (adjusted p < 0.001). A similar difference was also seen between the same patient groups in the right SMG and left SMG (adjusted p < 0.001). The RSFC of the right SMG and left SMG was also found to be significantly different between patient cohort AUD and NEITHER (adjusted p = 0.011).

Table 2.

Resting State Functional Connectivities Studied in Patient Groups.

RSFC: Combined RSFC: NEITHER RSFC: AUD RSFC: ANX RSFC: BOTH F(3,260) Hochberg p
ACC - L Alnsula −0.008 +/− 0.120 −0.011 +/− 0.015 0.008 +/− 0.015 −0.003 +/− 0.016 −0.028 +/− 0.014 1.00 0.64
ACC - R Alnsula −0.032 +/−0.134 −0.053 +/− 0.021 −0.012 +/− 0.014 −0.016 +/− 0.015 −0.048 +/− 0.016 1.66 0.21
ACC - L RPFC −0.035 +/− 0.131 −0.066 +/− 0.017 −0.018 +/− 0.014 −0.015 +/− 0.017 −0.042 +/− 0.016 2.21 0.10
ACC - R RPFC 0.079 +/−0.148 0.080 +/− 0.021 0.063 +/− 0.017 0.089 +/− 0.017 0.086 +/− 0.017 0.40 1.0
ACC - L SMG 0.136 +/− 0.136 0.144 +/− 0.020 0.142 +/− 0.016 0.124 +/− 0.016 0.133 +/− 0.015 0.31 1.0
ACC - R SMG 0.158 +/− 0.136 0.170 +/− 0.018 0.151 +/− 0.015 0.156 +/− 0.019 0.155 +/− 0.015 0.24 1.0
L Alnsula - R Alnsula 0.067 +/− 0.128 0.071 +/− 0.016 0.064 +/− 0.016 0.060 +/− 0.017 0.075 +/− 0.014 0.20 1.0
L Alnsula - L RPFC 0.069 +/− 0.136 0.082 +/− 0.021 0.063 +/− 0.015 0.058 +/− 0.018 0.072 +/− 0.012 0.40 1.0
L Alnsula - R RPFC 0.018 +/− 0.146 0.028 +/− 0.023 0.031 +/− 0.016 −0.006 +/− 0.018 0.018 +/− 0.014 0.84 0.90
L Alnsula - L SMG 0.365 +/− 0.148 0.405 +/− 0.017 0.352 +/− 0.018 0.347 +/− 0.021 0.356 +/− 0.017 2.20 0.10
L Alnsula - R SMG 0.275 +/− 0.168 0.317 +/− 0.027 0.261 +/− 0.017 0.260 +/− 0.018 0.260 +/− 0.018 1.87 0.16
R Alnsula - L RPFC −0.050 +/−0.130 −0.059 +/− 0.019 −0.053 +/− 0.015 −0.051 +/− 0.015 −0.038 +/− 0.015 0.31 1.0
R Alnsula - R RPFC −0.051 +/−0.112 −0.049 +/− 0.016 −0.060 +/− 0.013 −0.057 +/− 0.014 −0.038 +/− 0.013 0.50 1.0
R Alnsula - L SMG −0.043 +/−0.109 −0.053 +/− 0.015 −0.051 +/− 0.014 −0.040 +/− 0.012 −0.029 +/− 0.013 0.68 1.0
R Alnsula - R SMG −0.038 +/−0.125 −0.024 +/− 0.017 −0.049 +/− 0.015 −0.036 +/− 0.016 −0.045 +/− 0.014 0.49 1.0
L RPFC - R RPFC 0.023 +/− 0.121 0.040 +/− 0.018 0.002 +/− 0.012 0.020 +/− 0.014 0.031 +/− 0.015 1.18 0.42
L RPFC - L SMG 0.026 +/− 0.123 0.015 +/− 0.016 0.014 +/− 0.013 0.039 +/− 0.017 0.0350 +/− 0.014 0.76 1.0
L RPFC - R SMG 0.013 +/− 0.118 0.024 +/− 0.012 0.011 +/− 0.015 0.018 +/− 0.017 −0.002 +/− 0.014 0.56 1.0
R RPFC - L SMG 0.080 +/− 0.125 0.116 +/− 0.017 0.076 +/− 0.015 0.066 +/− 0.014 0.062 +/− 0.014 2.70 0.05
R RPFC-R SMG 0.085 +/− 0.128 0.132 +/− 0.019 0.086 +/− 0.013 0.077 +/− 0.015 0.044 +/− 0.014 5.73 <0.001
L SMG - R SMG −0.030 +/− 0.124 −0.081 +/− 0.018 −0.015 +/− 0.013 −0.032 +/− 0.014 0.006 +/− 0.015 6.21 <0.001
L Amy - ACC 0.055 +/− 0.170 0.048 +/− 0.023 0.084 +/− 0.022 0.048 +/− 0.019 0.041 +/− 0.019 0.85 0.84
L Amy - L Alnsula 0.026 +/− 0.173 0.042 +/− 0.024 0.038 +/− 0.020 −0.012 +/− 0.022 0.037 +/− 0.019 1.46 0.29
L Amy - R Alnsula 0.017 +/− 0.173 0.014 +/− 0.026 0.001 +/− 0.018 0.016 +/− 0.021 0.03 +/− 0.020 0.42 1.0
L Amy - L RPFC −0.021 +/− 0.174 −0.025 +/− 0.026 0.007 +/− 0.018 −0.024+/−0.020 −0.041 +/− 0.020 0.89 0.77
L Amy - R RPFC −0.010 +/− 0.179 0.011 +/− 0.026 0.007 +/− 0.022 −0.038 +/− 0.017 −0.021 +/− 0.022 1.10 0.50
L Amy - L SMG 0.014 +/− 0.167 0.018 +/− 0.024 0.015 +/− 0.020 0.002 +/− 0.018 0.020 +/− 0.021 0.15 1.0
L Amy - R SMG 0.029 +/− 0.182 0.025 +/− 0.027 0.030 +/− 0.021 0.020 +/− 0.022 0.040 +/− 0.020 0.15 1.0
R Amy - ACC −0.058 +/− 0.137 −0.053 +/− 0.018 −0.069 +/− 0.016 −0.046 +/− 0.017 −0.064 +/− 0.016 0.39 1.0
R Amy - L Alnsula −0.046 +/−0.143 −0.052 +/− 0.020 −0.057 +/− 0.016 −0.015 +/− 0.016 −0.062 +/− 0.018 1.51 0.26
R Amy - R Alnsula 0.002 +/− 0.161 −0.014+/−0.022 0.000 +/− 0.018 −0.010 +/− 0.021 0.031 +/− 0.018 1.05 0.56
R Amy - L RPFC 0.007 +/− 0.158 0.016 +/− 0.023 −0.006 +/− 0.018 −0.004+/−0.018 0.020 +/− 0.019 0.47 1.0
R Amy - R RPFC 0.019 +/− 0.131 0.026 +/− 0.014 0.022 +/− 0.016 0.013 +/− 0.017 0.017 +/− 0.017 0.12 1.0
R Amy - L SMG 0.068 +/− 0.142 0.063 +/− 0.018 0.094 +/− 0.016 0.0535 +/− 0.018 0.062 +/− 0.018 1.02 0.60
R Amy - R SMG 0.020 +/− 0.183 0.055 +/− 0.035 0.009 +/− 0.015 0.014 +/− 0.017 0.002 +/− 0.018 1.13 0.47
R Amy - L Amy 0.426 +/− 0.190 0.436 +/− 0.024 0.023 +/− 0.024 0.421 +/− 0.410 0.436 +/− 0.022 0.28 1.0
Open in a new tab

Fig. 2.

Fig. 2.

Open in a new tab

Bar graphs for the SN connectivities conducted with significant ANOVA p-values are displayed. Significant results of post-hoc analysis via Tukey’s HSD displayed. A) R SMG to R RPFC: BOTH vs NEITHER (p < 0.001). B) R SMG to L SMG: BOTH vs NEITHER (p < 0.001) and AUD vs NEITHER (p = 0.011).

Two-way ANOVAs were performed to assess the interactions between AUD and ANX on RSFC between these regions, evaluating whether their effects were synergistic or additive. For the right SMG and right RPFC, the analyses revealed significant main effects for AUD (F (1,260) = 6.72; p = 0.010) and ANX (F(1,260) = 10.30; p = 0.0015), but no significant interaction effect (F(1,260) = 0.18; p = 0.67). Similarly, for the connectivity between the right SMG and left SMG, significant main effects were observed for AUD (F(1,260) = 12.20; p < 0.001) and ANX (F(1,260) = 5.55; p = 0.019), with no significant interaction effect (F(1,260) = 0.87; p = 0.35). This lack of interaction suggests that the effects of anxiety and AUD on RSFC may be additive rather than synergistic. However, it’s important to note that the absence of a significant interaction does not conclusively confirm additivity; it merely indicates that the combined effect does not significantly differ from the sum of the individual effects in our sample.

Finally, we correlated dimensional measures of alcohol use (WHO ASSIST) and anxiety (GAD-7) against both individual patient groups (Fig. 3A, B, 4A, B) and the combined cohort for each connectivity of interest (Fig. 3C, 4C). Exploratory correlations revealed two nominally significant correlations between clinical scores and RSFC: WHO ASSIST scores negatively correlated with right SMG to right RPFC RSFC (r = 0.14; p = 0.02 - Fig. 3C) and positively correlated with interhemispheric SMG RSFC (r = 0.17; p = 0.006 - Fig. 3C).

Fig. 3.

Fig. 3.

Open in a new tab

A) Correlations between WHO ASSIST scores and RSFC of Right SMG to Right RPFC within each patient group. B) Correlations between WHO ASSIST scores and RSFC of Right SMG to Left SMG within each patient group. C) Correlations between WHO ASSIST scores and RSFC for both connectivities with all four groups merged. Asterisks indicate statistical significance: p < 0.05 (*), p < 0.01 (**).

Fig. 4.

Fig. 4.

Open in a new tab

A) Correlations between GAD-7 scores and RSFC of Right SMG to Right RPFC within each patient group. B) Correlations between GAD-7 scores and RSFC of Right SMG to Left SMG within each patient group. C) Correlations between GAD-7 scores and RSFC for both connectivities.

4. Discussion

This study investigated differences in RSFC within the SN and to the amygdala among psychiatric inpatients diagnosed with alcohol use disorder (AUD), anxiety disorders (ANX), both (BOTH), or neither (NEITHER). Using fMRI and connectivity analysis, we identified significant RSFC differences, particularly between the right SMG and 1) the right RPFC and 2) the left SMG. These connectivity differences may indicate disrupted integration of cognitive and emotional processing in these patients. Specifically, post hoc analysis revealed significantly higher RSFC in the NEITHER cohort compared to the BOTH cohort in right SMG to right RPFC connectivity. This suggests that patients with both AUD and ANX exhibit markedly different connectivity patterns compared to those without either condition. Significant differences in interhemispheric SMG connectivity between the BOTH and NEITHER cohorts, as well as between the BOTH and AUD cohorts, with connectivity highest in the BOTH cohort, followed by AUD and NEITHER, underscore the compounded impact of comorbid anxiety and alcohol use on brain connectivity.

Exploratory correlations between RSFC and clinical measures of alcohol use (WHO ASSIST) and anxiety (GAD-7) suggest potential links between the severity of clinical symptoms and neural connectivity. Specifically, interhemispheric RSFC of the SMG is significantly positively correlated with higher severity of alcohol use and anxiety RSFC between the right SMG and the right RPFC is significantly negatively correlated with alcohol use. The observed trends suggest that RSFC may hold promise as an indicator of disease severity and progression in the context of alcohol use, warranting further investigation. These findings align with prior research suggesting that functional connectivity measures can serve as indicators of clinical states and inform treatment strategies (Grodin et al., 2019).

As expected, the WHO ASSIST scores consistently showed a disparity of approximately 20 points (on a 39-point scale) between the AUD/BOTH groups and the NEITHER/ANX groups. This outcome was expected, as alcohol use is a quantifiable trait integral to diagnostic criteria. However, the GAD-7 scores did not exhibit a similar difference. This suggests that anxiety symptoms may be pervasive across the entire psychiatric inpatient population, regardless of DSM diagnosis. This finding could indicate that severe anxiety is also widespread among psychiatric inpatients and that it transcends traditional diagnostic boundaries. This may prompt a reconsideration of how we conceptualize and address anxiety within this population.

Our findings underscore the importance of considering comorbidity when trying to understand the neural mechanisms underlying these psychiatric conditions. The presence of both disorders appears to exacerbate disruptions in neural connectivity within key regions associated with cognitive and emotional regulation. Disruption in SN activity may cause an excessive focus on negative or threatening stimuli, thereby heightening the risk of developing and sustaining AUD and anxiety disorders (Schimmelpfennig et al., 2023). Research continues to explore the SN’s comprehensive role in these complex conditions, aiming to better understand its potential as a therapeutic target. This includes investigating how modulation of this network could help alleviate symptoms or even prevent the progression of these disorders (Padula et al., 2022; Manza et al., 2023). Since targeting the SN by modulating its activity through transcranial magnetic stimulation has been shown to possibly be effective in treating AUD suggesting a promising area for therapeutic interventions (Padula et al., 2022), it is possible that in patients with comorbid AUD-ANX this approach may be particularly efficient: Our data points to an additive effect of both conditions on RSFC. Our sample consisted of four cohorts of patients, matched on demographic and clinical characteristics, which allowed comparisons between cohorts rather than against healthy controls. Research in substance use (and psychiatry in general) often includes healthy subjects as controls, with groups commonly matched for demographic factors only. However, in those cases, even when healthy controls are carefully recruited to match the patients, the results could stem from various other disorders or symptoms rather than being specific to the studied disorder, due to high comorbidity with other psychiatric disorders (Gosnell et al., 2020). Most psychiatric inpatients share several characteristics, such as chronic stress or poor sleep, which, being absent in the control group, could cause any of the observed results. In our study, we matched both demographic characteristics and psychiatric diagnoses between patient cohorts. The use of inpatient data, from individuals with comorbid psychiatric conditions is a strength of this study. This approach enhances the ecological validity of the findings, allowing them to be more directly applicable to inpatient populations where such comorbidities are common. Consequently, we are more confident that AUD or ANX, and not common comorbidities, were the variables that discriminated RSFC within the SN.

Discrepancies in how the salience network is defined and categorized underscore the complexity of investigating large-scale brain systems (Uddin et al., 2019). These inconsistencies often result in overlapping terms and frameworks for describing the same network (Yeo et al., 2015). In this study, we addressed these challenges by utilizing the prebuilt ROI definitions provided by the CONN toolbox. This approach ensured that our analysis was grounded in a standardized system, reducing ambiguity in ROI selection while maintaining compatibility with existing literature.

Although our study did not yield significant results regarding the amygdala, existing literature highlights its critical role in emotional processing, including in the context of anxiety and substance use disorders (Anker, 2019; Smith et al., 2021). An explanation for the lack of significant results in the amygdala may be our use of psychiatric patients as controls. Much of the amygdala-related findings in the literature have been derived from comparisons with healthy controls, which could render those results more influenced by general psychiatric comorbidities (Gosnell et al., 2020). An additional consideration is the use of a single amygdala ROI, which averaged time series across functionally distinct subregions. This approach could dilute meaningful connectivity patterns, particularly in areas of low BOLD signal intensity. Further research into the amygdala’s involvement could offer valuable insights into the neurobiological mechanisms underlying these comorbid conditions and guide the development of targeted treatments.

While this study provides valuable insights, its findings are subject to several limitations. First, the resting state scans were only 5 min long, although the time required for reliable estimation of functional connectivity is now known to be larger (Tomasi et al., 2017). Second, while our ANOVA results were robust and survived Hochberg correction, the correlations between RSFC and clinical measures were not subject to the same correction, suggesting that the observed correlations should be interpreted with caution. Third, while our sample size was adequate, the possibility remains that some significant results were not detected due to limited statistical power. However, the strategy we utilized—hypothesis-supported ROI to ROI with Hochberg correction—is statistically more robust than ‘seed-to-voxel,’ where noise, by design, can falsely identify significant results (Lyon, 2017). Fourth, there is a lack of information regarding substance use frequency in our sample. However, the WHO ASSIST measure provides a more holistic metric by capturing the lifestyle effects of substance use rather than focusing solely on amount or frequency. Fifth, our results may reflect effects of detoxification treatment for AUD rather than an underlying neurobiological deficit, which could limit the generalizability outside inpatient settings. However, this context enhances the relevance of our findings for inpatient populations, where understanding neurobiological changes during detoxification can provide critical insights to inform treatment strategies and improve patient outcomes. Future studies should aim to compare RSFC in patients before and after detoxification. Finally, our sample lacked ethnic and socioeconomic diversity as at the time, most inpatients at the Menninger Clinic were of white ancestry and affluent. This potentially limits the generalizability of our findings to broader populations, although it also reduces its effect as a confounding variable. Future research should aim to replicate these results in larger, more diverse cohorts to confirm their validity and robustness. Additionally, longitudinal studies could provide valuable insights into whether changes in RSFC precede or follow clinical symptomatology, thereby informing prevention and intervention strategies.

5. Conclusion

In summary, we have demonstrated significant alterations in RSFC within the salience network (SN) among psychiatric inpatients with alcohol use disorder (AUD) and anxiety disorders (ANX). Our findings revealed lower functional connectivity between the right supramarginal gyrus (SMG) and the rostral prefrontal cortex (RPFC) in patients with both AUD and ANX compared to those with neither condition, suggesting a further disrupted integration of cognitive and emotional processing when both conditions are present. The interhemispheric connectivity differences within the SMG further indicate potential disruptions in brain communication, also exacerbated by the comorbidity of AUD and ANX. These differential connectivity patterns highlight the RPFC and SMG as key regions in the neural architecture underlying these conditions, pointing to likely additive effects. Our data suggest that the SN and its associated regions could be promising targets for neuromodulation or psychopharmacological interventions, which could inform the development of improved therapeutic strategies for AUD and ANX comorbidity.

Acknowledgements

The authors thank the Core for Advanced MRI (CAMRI) at Baylor College of Medicine, Dr. Charles Neblett, and participants and their families. Funding: This study was supported by the Veterans Health Administration (grant VHA I01CX001937 to RS) and The Robert & Janice McNair Foundation. This study is in part the result of work supported with resources and the use of facilities at the Michael E. DeBakey Veterans Affairs Medical Center. These data included herein were collected through the use of facilities and resources at The Menninger Clinic, Houston, Texas USA. The content is solely the responsibility of the authors and does not represent the official views of the Department of Veterans Affairs or the United States government.

Funding

Funders had no role in the design, data collection, analysis, or writing of the paper.

Footnotes

CRediT authorship contribution statement

Dhruv M. Patel: Writing – review & editing, Writing – original draft, Formal analysis, Conceptualization. Guillermo Poblete: Writing – review & editing, Formal analysis, Data curation. Alexandra Castellanos: Writing – review & editing, Project administration, Data curation. Ramiro Salas: Writing – review & editing, Writing – original draft, Supervision, Project administration, Funding acquisition, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare no conflict of interest.

References

  1. Ambrosi E et al. , “Resting-state functional connectivity of the Habenula in mood disorder patients with and without suicide-related behaviors,” J. Neuropsychiatr. Clin. Neurosci, vol. 31, 49, Winter 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ambrosi E et al. , “Resting-state functional connectivity of the anterior cingulate cortex among persons with mood disorders and suicidal behaviors,” J. Neuropsychiatr. Clin. Neurosci, vol. 36, 143, Spring 2024. [DOI] [PubMed] [Google Scholar]
  3. Anker JJ et al. , “Co-Occurring Alcohol Use Disorder and Anxiety: Bridging Psychiatric, Psychological, and Neurobiological Perspectives,” Alcohol Res, vol. 40, 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Behzadi Y, et al. , 2007. A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage 37 (1), 90. Aug. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bush G, et al. , 2000. Cognitive and emotional influences in anterior cingulate cortex. Trends Cogn. Sci 4, 215. Jun. [DOI] [PubMed] [Google Scholar]
  6. Bzdok D, et al. , 2013. Segregation of the human medial prefrontal cortex in social cognition. Front. Hum. Neurosci 7, 232. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Conway KP, et al. , 2006. Lifetime comorbidity of DSM-IV mood and anxiety disorders and specific drug use disorders: results from the National Epidemiologic Survey on alcohol and related conditions. J. Clin. Psychiatry 67, 247. Feb. [DOI] [PubMed] [Google Scholar]
  8. Curtis K, et al. , 2017. Increased habenular connectivity in opioid users is associated with an alpha5 subunit nicotinic receptor genetic variant. Am. J. Addict 26, 751. Oct. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Curtis K, et al. , 2019. Insular resting state functional connectivity is associated with gut microbiota diversity. Eur. J. Neurosci 50, 2446. Aug. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cushnie AK, et al. , 2023. Connecting circuits with networks in addiction neuroscience: a salience network perspective. Int. J. Mol. Sci 24. May 22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dosenbach NU, et al. , 2007. Distinct brain networks for adaptive and stable task control in humans. Proc. Natl. Acad. Sci. USA 104 (26), 11073. Jun. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Driessen M, et al. , 2001. The course of anxiety, depression and drinking behaviours after completed detoxification in alcoholics with and without comorbid anxiety and depressive disorders. Alcohol Alcohol. 36, 249. May-Jun. [DOI] [PubMed] [Google Scholar]
  13. First MB, et al. , 2002. Structured Clinical Interview for DSM-IV-TR Axis I Disorders. American Psychiatric Publications. [Google Scholar]
  14. Fox MD, et al. , 2006. Spontaneous neuronal activity distinguishes human dorsal and ventral attention systems. Proc. Natl. Acad. Sci. USA 103, 10046. Jun 27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fukushima S, et al. , 2020. Behavioural cue reactivity to alcohol-related and non-alcohol-related stimuli among individuals with alcohol use disorder: an fMRI study with a visual task. PLoS One 15, e0229187. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gilbert SJ, et al. , 2006. Functional specialization within rostral prefrontal cortex (area 10): a meta-analysis. J. Cogn. Neurosci 18, 932. Jun. [DOI] [PubMed] [Google Scholar]
  17. Gordon EM, et al. , 2017. Precision functional mapping of individual human brains. Neuron 95, 791. Aug 16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gorka SM, et al. , Sep 2013. Alcohol attenuates amygdala-frontal connectivity during processing social signals in heavy social drinkers: a preliminary pharmaco-fMRI study. Psychopharmacology (Berl) 229, 141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gosnell SN, et al. , 2020. Hippocampal Volume in Psychiatric Diagnoses: Should Psychiatry Biomarker Research Account for Comorbidities? Chronic Stress (Thousand Oaks) 4, 2470547020906799, Jan-Dec. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Greene DJ, et al. , 2016. Considerations for MRI study design and implementation in pediatric and clinical populations. Dev. Cogn. Neurosci 18, 101. Apr. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Grodin EN, et al. , 2019. The use of functional magnetic resonance imaging to test pharmacotherapies for alcohol use disorder: a systematic review. Alcohol. Clin. Exp. Res 43, 2038. Oct. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Humeniuk R, et al. , 2008. Validation of the alcohol, smoking and substance involvement screening test (ASSIST). Addiction 103, 1039. Jun. [DOI] [PubMed] [Google Scholar]
  23. Khantzian EJ, 1997. The self-medication hypothesis of substance use disorders: a reconsideration and recent applications. Harv. Rev. Psychiatry 4, 231. Jan-Feb. [DOI] [PubMed] [Google Scholar]
  24. Khoury NM, et al. , 2018. Interoception in psychiatric disorders: a review of randomized, controlled trials with Interoception-based interventions. Harv. Rev. Psychiatry 26, 250. Sep/Oct. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kushner MG, et al. , 1999. Prospective analysis of the relation between DSM-III anxiety disorders and alcohol use disorders. Am. J. Psychiatry 156, 723. May. [DOI] [PubMed] [Google Scholar]
  26. Kushner MG, et al. , 2000. The relationship between anxiety disorders and alcohol use disorders: a review of major perspectives and findings. Clin. Psychol. Rev 20, 149. Mar. [DOI] [PubMed] [Google Scholar]
  27. Kushner MG, et al. , 2005. Follow-up study of anxiety disorder and alcohol dependence in comorbid alcoholism treatment patients. Alcohol. Clin. Exp. Res 29, 1432. Aug. [DOI] [PubMed] [Google Scholar]
  28. LaBounty LP, et al. , 1992. Relapse among alcoholics with phobic and panic symptoms. Addict. Behav 17, 9. [DOI] [PubMed] [Google Scholar]
  29. Liu WJ, et al. , 2015. Abnormal functional connectivity of the amygdala-based network in resting-state FMRI in adolescents with generalized anxiety disorder. Med. Sci. Monit 21, 459. Feb 12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lyon L, 2017. Dead salmon and voodoo correlations: should we be sceptical about functional MRI? Brain 140, e53. Aug 1. [DOI] [PubMed] [Google Scholar]
  31. Manza P, et al. , 2023. Neural circuit selective for fast but not slow dopamine increases in drug reward. Nat. Commun 14, 6408. Nov 8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Menon V, et al. , 2010. Saliency, switching, attention and control: a network model of insula function. Brain Struct. Funct 214, 655. Jun. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Moss AC, et al. , 2015. The effects of responsible drinking messages on attentional allocation and drinking behaviour. Addict. Behav 44, 94. May. [DOI] [PubMed] [Google Scholar]
  34. Munjuluri S, et al. , 2020. A pilot study on playback theatre as a therapeutic aid after natural disasters: brain connectivity mechanisms of effects on anxiety. Chronic Stress (Thousand Oaks) 4, 2470547020966561, Jan-Dec. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Naqvi NH, et al. , 2010. The insula and drug addiction: an interoceptive view of pleasure, urges, and decision-making. Brain Struct. Funct 214, 435. Jun. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Padula CB, et al. , 2022. Targeting the salience network: a Mini-review on a novel neuromodulation approach for treating alcohol use disorder. Front. Psychol 13, 893833. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Paulus MP, et al. , 2005. Dose-dependent decrease of activation in bilateral amygdala and insula by lorazepam during emotion processing. Arch. Gen. Psychiatry 62, 282. Mar. [DOI] [PubMed] [Google Scholar]
  38. Peters SK, et al. , 2016. Cortico-striatal-thalamic loop circuits of the salience network: a central pathway in psychiatric disease and treatment. Front. Syst. Neurosci 10, 104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Plummer F, et al. , 2016. Screening for anxiety disorders with the GAD-7 and GAD-2: a systematic review and diagnostic metaanalysis. Gen. Hosp. Psychiatry 39, 24. Mar-Apr. [DOI] [PubMed] [Google Scholar]
  40. Poblete G, et al. , 2022. A novel approach to link genetics and human MRI identifies AKAP7-dependent Subicular/prefrontal functional connectivity as altered in suicidality. Chronic Stress (Thousand Oaks) 6, 24705470221083700, Jan-Dec. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Rufino KA, et al. , 2024. Val/met BDNF as a genetic risk for a false sense of security in post-discharge suicide risk. J. Affect. Disord 354, 98. Jun 1. [DOI] [PubMed] [Google Scholar]
  42. Schimmelpfennig J, et al. , 2023. The role of the salience network in cognitive and affective deficits. Front. Hum. Neurosci 17, 1133367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Seeley WW, et al. , 2007. Dissociable intrinsic connectivity networks for salience processing and executive control. J. Neurosci 27, 2349. Feb 28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Shin LM, et al. , 2010. The neurocircuitry of fear, stress, and anxiety disorders. Neuropsychopharmacology 35, 169. Jan. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Silani G, et al. , 2013. Right supramarginal gyrus is crucial to overcome emotional egocentricity bias in social judgments. J. Neurosci 33, 15466. Sep 25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Simmons A, et al. , 2006. Anticipation of aversive visual stimuli is associated with increased insula activation in anxiety-prone subjects. Biol. Psychiatry 60, 402. Aug 15. [DOI] [PubMed] [Google Scholar]
  47. Smith DM, et al. , 2021. Valence encoding in the amygdala influences motivated behavior. Behav. Brain Res 411, 113370. Aug 6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Tomasi DG, et al. , 2017. Temporal evolution of brain functional connectivity metrics: could 7 min of rest be enough? Cereb. Cortex 27, 4153. Aug 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Tzourio-Mazoyer N, et al. , 2002. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. Neuroimage 15, 273. Jan. [DOI] [PubMed] [Google Scholar]
  50. Uddin LQ, 2015. Salience processing and insular cortical function and dysfunction. Nat. Rev. Neurosci 16, 55. Jan. [DOI] [PubMed] [Google Scholar]
  51. Uddin LQ, 2016. Salience Network of the Human Brain. Academic Press is an imprint of Elsevier, London, United Kingdom, p. x. [Google Scholar]
  52. Uddin LQ, et al. , 2019. Towards a universal taxonomy of macro-scale functional human brain networks. Brain Topogr. 32, 926. Nov. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Vickerstaff V, et al. , 2019. Methods to adjust for multiple comparisons in the analysis and sample size calculation of randomised controlled trials with multiple primary outcomes. BMC Med. Res. Methodol 19, 129. Jun 21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Whitfield-Gabrieli S, et al. , 2012. Conn: a functional connectivity toolbox for correlated and anticorrelated brain networks. Brain Connect. 2, 125. [DOI] [PubMed] [Google Scholar]
  55. Wilmer MT, et al. , 2021. Correlates of quality of life in anxiety disorders: review of recent research. Curr. Psychiatry Rep 23, 77. Oct 6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Won E, et al. , 2020. Neuroinflammation-associated alterations of the brain as potential neural biomarkers in anxiety disorders. Int. J. Mol. Sci 21. Sep 7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Wrase J, et al. , 2008. Amygdala volume associated with alcohol abuse relapse and craving. Am. J. Psychiatry 165, 1179. Sep. [DOI] [PubMed] [Google Scholar]
  58. Yang W, et al. , 2022. Alcohol use disorder: neurobiology and therapeutics. Biomedicines 10. May 21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Yeo BT, et al. , 2015. Functional specialization and flexibility in human association cortex. Cereb. Cortex 25, 3654. Oct. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES